Prothena Corporation’s partners Roche and Bristol Myers Squibb presented updated clinical data at the AD/PD 2026 international conference in Copenhagen, covering two investigational antibodies targeting distinct protein aggregation pathways: prasinezumab, directed at aggregated alpha-synuclein in Parkinson’s disease, and BMS-986446, an anti-tau antibody in early-stage development for Alzheimer’s disease. The presentations, spanning oral sessions and posters delivered between 17 and 21 March 2026, consolidate a body of evidence that is encouraging for both programmes while stopping short of the definitive efficacy proof the field requires.
Why the ‘two years saved’ framing for prasinezumab matters to regulators and clinicians
The most analytically significant prasinezumab presentation at AD/PD 2026 was not a conventional trial readout but a modelling exercise. Roche presented a pharmacometric analysis comparing outcomes from the PASADENA open-label extension against predictions derived from the Parkinson’s Progression Markers Initiative, a large independent observational cohort. The model estimated that participants in the PASADENA trial were, on average, approximately two years less advanced in disease severity at the five-year mark than would be expected without treatment. The framing of this as ‘time saved’ is deliberate and consequential: regulatory bodies have historically struggled to accept disease modification claims in Parkinson’s disease because motor rating scales are confounded by symptomatic medications. A time-based efficacy metric, if validated, offers a more intuitive and clinician-friendly measure of benefit that could form the basis of a labelling argument.
However, the modelling approach carries inherent limitations that the field will scrutinise carefully. External comparator arms derived from observational registries, such as PPMI, are subject to selection bias, different assessment frequencies, and site-level variability that randomised controlled trials are designed to eliminate. Regulatory watchers note that while such analyses can reinforce a signal hypothesis, they are unlikely to serve as primary evidence for approval without corroboration from a prospectively powered Phase III trial with pre-specified endpoints.
What the PADOVA open-label extension adds beyond the missed primary endpoint
The topline PADOVA Phase IIb result, reported in December 2024, missed statistical significance on its primary endpoint of time to confirmed motor progression, with a hazard ratio of 0.84 and a p-value of 0.0657. The AD/PD 2026 presentations drew on longer-term open-label extension data from PADOVA, showing a sustained slowing of motor decline on top of levodopa and monoamine oxidase-B inhibitor therapy. While this direction of effect is consistent, the absence of a concurrent placebo arm in the extension phase means confounding cannot be excluded. The signal is nonetheless meaningful because PADOVA enrolled participants already on effective symptomatic treatment, a population in which detecting an additive disease-modifying effect is substantially harder than in treatment-naive patients.
A pre-specified subgroup analysis of PADOVA had already shown a more pronounced effect in the levodopa-treated population, with a hazard ratio of 0.79 that did achieve nominal significance. This subgroup finding is consistent across two trials and has now been reinforced by exploratory biomarker data from imaging, which showed a numerically slower loss of neuromelanin signal in the substantia nigra pars compacta and a statistically significant reduction in iron accumulation in the putamen. Taken together, these results suggest biological activity at the target tissue level, an important distinction from trials where clinical effect alone drives the argument. Whether the levodopa interaction represents a true pharmacological synergy or a population enrichment artefact remains an open and important question for the Phase III design.
How digital health endpoints are reshaping what counts as evidence in Parkinson’s trials
A poster at AD/PD 2026 described post-hoc analyses of digital health technology data collected during PADOVA in the practically-defined OFF L-dopa state, a window when symptomatic drug effects are minimised and underlying disease severity is most visible. The findings showed consistent trends favouring prasinezumab across digitally captured motor parameters, aligning with earlier findings from the PASADENA Phase IIa simple sum digital endpoint. This convergence across two trials and two different data collection methodologies strengthens the biological signal, even if neither dataset is powered to serve as a standalone efficacy measure.
The inclusion of digital endpoints in both trials reflects a broader methodological shift in Parkinson’s disease research. Regulators and trial designers have grown increasingly interested in continuous monitoring data as a complement to periodic clinical assessments, given that the latter capture only a narrow slice of patient function. Industry observers note that if digital metrics continue to track with clinical outcomes in the Phase III PARAISO study, they could eventually form part of the evidence base for a disease-modification claim, though formal regulatory acceptance of such endpoints as primary measures remains to be established.
What Roche’s Phase III commitment reveals about its confidence in a contested dataset
Roche announced its decision to initiate the Phase III PARAISO study in June 2025, explicitly citing the totality of evidence from PASADENA and PADOVA rather than any single statistically significant result. This is a meaningful commercial and scientific commitment: a Phase III Parkinson’s disease trial enrolling a broad early-stage population on standard-of-care therapy will require substantial resources, extended timelines, and a robust endpoint strategy. The decision signals that Roche’s internal evidence threshold for proceeding was met even in the absence of a formal statistical win, a posture that will invite scrutiny from the investment community and health technology assessment bodies who will eventually evaluate cost-effectiveness.
The alpha-synuclein targeting rationale itself remains scientifically credible but has not yet produced a therapy with regulatory approval. Prasinezumab is designed to bind aggregated forms of alpha-synuclein and reduce cell-to-cell spread, a mechanism grounded in post-mortem and preclinical evidence. The absence of a validated CSF or plasma biomarker for aggregated alpha-synuclein that correlates reliably with clinical outcomes makes it harder to confirm target engagement in vivo and adds interpretive complexity to every dataset. The neuromelanin and iron MRI data presented at AD/PD 2026 represent a step toward biomarker validation, but neither measure is yet an accepted surrogate endpoint for regulatory purposes.
BMS-986446 safety profile in Japanese participants and what it changes for the Phase 2 trial
The Bristol Myers Squibb presentation at AD/PD 2026 covered Phase 1 single-dose safety and pharmacokinetics data for BMS-986446, with specific attention to Japanese participants. The study confirmed dose-proportional plasma exposure, an absence of anti-drug antibodies, and a safety profile that was consistent across ethnicities, supporting the use of uniform dosing in the ongoing Phase 2 TargetTau-1 trial without ethnic-specific adjustments. While a clean Phase 1 safety readout is expected rather than surprising for a monoclonal antibody with a well-characterised mechanism, the explicit Japanese cohort data carries practical importance: Japan represents a significant Alzheimer’s disease burden and regulatory market, and early ethnicity bridging studies reduce the likelihood of needing a separate regional development programme.
BMS-986446 targets the R1 to R3 repeats within the microtubule binding region of tau, a mechanistic focus that distinguishes it from prior tau antibodies that attacked the N-terminal or C-terminal regions of the protein. Those earlier approaches, including semorinemab (Roche) and gosuranemab (Biogen), failed to demonstrate clinical efficacy in Phase 2 trials. The MTBR rationale, supported by structural biology showing that this domain drives seeded aggregation, represents the field’s current best hypothesis for where antibody intervention can disrupt tau spread. The FDA granted TargetTau-1 Fast Track Designation in October 2025, acknowledging both the unmet need and the mechanistic plausibility of the approach.
Why the MTBR tau hypothesis carries more conviction than its predecessors and what could still go wrong
The shift from terminal-region to mid-region tau targeting reflects a hard-won lesson from a decade of clinical failures. Industry observers note that MTBR fragments are the forms of tau most strongly associated with neurofibrillary tangle formation, and that preclinical models have repeatedly shown MTBR antibodies to be superior to terminal-region alternatives at blocking both tau uptake into neurons and cell-to-cell propagation. BMS-986446 also engages Fc receptor-mediated microglial phagocytosis, adding an active clearance mechanism on top of the propagation-blocking effect. This dual mode of action is a genuine differentiator relative to most previous tau antibodies.
The critical unknowns remain the same for any tau programme. First, by the time patients meet the clinical criteria for early Alzheimer’s disease, substantial tau pathology has already accumulated, raising the question of whether halting further spread can translate into meaningful cognitive stabilisation. Second, CSF antibody concentrations typically reach approximately 0.2 percent of plasma levels for conventional monoclonal antibodies, a figure confirmed in BMS-986446’s Phase 1 data. Whether this penetration is sufficient to engage MTBR tau throughout the brain parenchyma at therapeutically relevant concentrations is an unresolved question that the Phase 2 tau-PET endpoints in TargetTau-1 will need to address. Third, BMS-986446 is entering a competitive field that now includes established amyloid-targeting agents with accelerated or full approval, meaning that any tau programme seeking reimbursement will need to demonstrate additive or complementary value beyond what is already available.
How Prothena’s partnered pipeline model shapes its risk exposure across both programmes
Prothena’s position in both the prasinezumab and BMS-986446 programmes is that of a royalty and milestone beneficiary rather than an active development partner. Under its agreement with Bristol Myers Squibb, Prothena is eligible to receive up to USD 562.5 million in additional regulatory and sales milestone payments, plus tiered royalties on net sales, with the Swiss pharmaceutical group assuming responsibility for all development, manufacturing, and commercial activities. This structure limits Prothena’s direct capital exposure but also means the Dublin-based biotech has limited control over design decisions in either the TargetTau-1 Phase 2 trial or the forthcoming Phase III PARAISO study. For investors and analysts tracking Prothena’s pipeline optionality, the concentration of value in Roche and Bristol Myers Squibb execution creates a binary-like risk profile that is not fully captured by clinical milestone announcements alone.
What clinicians, regulators, and pipeline watchers should monitor next
The immediate priorities for both programmes are well-defined. For prasinezumab, the PARAISO Phase III trial design, specifically its primary endpoint choice, patient population definition, and the regulatory agreement on whether a biomarker-enriched subgroup can serve as the primary analysis, will be the determining factor in whether a decade of alpha-synuclein research ultimately yields an approved therapy. For BMS-986446, the TargetTau-1 Phase 2 results, expected no earlier than 2027 given the trial’s completion timeline, will need to demonstrate not only safety but a measurable reduction in tau-PET signal and a clinical benefit signal on the Clinical Dementia Rating Sum of Boxes. Regulatory watchers will also be watching whether the tau-PET endpoint can be positioned as a surrogate, as amyloid-PET was for the approved amyloid antibodies, to enable an accelerated pathway.
Across both programmes, the AD/PD 2026 presentations represent evidence consolidation rather than evidence generation. The signal is consistent and the mechanistic rationale credible, but neither prasinezumab nor BMS-986446 has yet produced the kind of prospectively powered, statistically clean efficacy data that regulators require for approval. The translation from protein biology to clinical benefit in neurodegenerative disease remains the field’s central unsolved problem, and both programmes will be judged ultimately on whether they can close that gap.
Why Roche’s Phase III bet on prasinezumab is bigger than its Phase IIb result suggests added by Pallavi Madhiraju on
View all posts by Pallavi Madhiraju →